1. Field of the Invention (Technical Field)
The present invention relates to a method of improving the uniformity of standardized microwave exposures while maximizing the exposure field level for a given Radio Frequency (RF) input power. The present invention is an extension of previous methods in that it is applicable to devices and materials that exhibit frequency-selective microwave absorption, actively adjusting the shielded chamber excitation to maintain a uniform average microwave exposure field.
2. Description of Related Art
RF microwave exposures can be performed in many ways. It can be performed in the free field, often the most realistic situation. Of course, the escape of microwave radiation to the surroundings is guaranteed. Furthermore, high power sources are often needed to assure sufficient electric field at a Device Under Exposure (DUE). Anechoic chambers are used to obviate the first problem: the microwave radiation is contained and absorbed by special wall linings inside of a shielded cavity. The problem still exists that high power sources are necessary for a wide range of exposures, that the direction of incidence of the microwave radiation is from one direction at a time only, and depending on the source, exposures are performed in one electric polarization at a time. RF testing or industrial microwave exposures then become an onerous procedure, onerous especially if only high-level, global results are required.
Reverberation chambers lack the absorbing wall linings that anechoic chamber incorporate. The microwave radiation “fills” the shielded cavity, bouncing back and forth between the walls. The multiplication of the electric field strength in the cavity “hot spots” can be enormous: this is how a microwave oven works. Much lower power sources are required. Furthermore, testing is performed at all angles, directions of incidence, and all polarizations as the radiation bounces around the cavity. The benefits are large for production line microwave exposures: the weakness, however, is equally large. The weakness associated with reverberation chambers is that they have electric cold spots and hot spots interspersed throughout the volume, depending on how the electric fields add and cancel as they bounce around. The position of these can dramatically affect the electric field level to which the DUE is exposed. Worse, the positions of the hot and cold spots depend on the test frequency, as well as the location and orientation of the DUE.
Two solutions have been implemented to mitigate these disadvantages in reverberation chambers. The traditional solution is the one that has received the greatest investment in development; it is known as “mechanical mode-stir”. By incorporating a significant moving mechanical feature inside the chamber, the positions of the hot-spots and cold-spots can be shifted back and forth, up and down, so that on average the electric field is the same everywhere inside the chamber. Typical mode-stir devices are giant fans, whose blades rotate to change the effective boundary location of the wall, and rotate rapidly in an attempt to minimize the time that the electric field hot spots exist in any one location. The average shifting times in a good mechanical mode-stir equipped device are in the milliseconds. Despite the fact that modern electronics respond to time scales up to one million times faster than this, mechanical mode-stir has been developed to provide a reliable, well-defined and reasonably inexpensive method of testing.
The competing technology in reverberation chamber microwave exposures is electronic mode-stir. With this background, electronic mode-stir is simple to qualitatively define. Since the positions of the chamber hot-spots and cold-spots depend on the excitation frequency, then by rapidly changing the excitation frequency over a range of values, then these hot-spots and cold-spots can essentially “blur” together. This is accomplished in practice by either sweeping the excitation up or down in frequency, or, by randomly shifting from one frequency or phase to another. The latter method is often referred to as noise excitation.
Noise excitation is particularly simple to implement. A noise signal is combined with a microwave center frequency signal at a balanced mixer, and the result is a fairly uniform band of random frequencies generated above and below that center frequency. With clever system design, this results in a reverberation chamber having an average electric field value everywhere in the chamber, regardless of the center frequency or of the position of the DUE in the chamber.
It is informative to define electronic mode-stir more quantitatively. For a rectangular cavity with perfectly conducting walls, the frequencies at which resonances can exist is given by equation 1, wherein (l, m, n) is any set of non-negative integers, except (0, 0, 0), (a, b, d) are the dimensions of the rectangular cavity, and c is the speed of light.
                              f          lmn                =                              c            2                    ⁢                                                                                          (                                          l                      a                                        )                                    2                                +                                                      (                                          m                      b                                        )                                    2                                +                                                      (                                          n                      d                                        )                                    2                                                      .                                              Equation        ⁢                                  ⁢        1            
Each mode of the chamber then can be designated by the integer triple (l, m, n). If a is the largest chamber dimension (they may all be equal), then the base frequency of the rectangular chamber corresponds to the mode (1, 0, 0). Below this the chamber is electrically small compared to the excitation frequency, and cannot sustain any resonance at all. As l, m, and n increase and became large, then the frequencies flmn become closer together and eventually form a nearly continuous variable. The result is often called an “over-moded” chamber. The nearly continuous spectrum of chamber excitation frequencies is denoted simply by ‘f’.
At large l, m, and n, the total number of all modes excited up to such a high frequency, N, can be closely approximated by Equation 2:N≈lmn  Equation 2.In this high-frequency approximation, Equation 3 shows that for some constant factor γ, the number of modes ΔN within some frequency interval Δf centered at a frequency f is:
                              Δ          ⁢                                          ⁢          N                =                                            24              ⁢              V                                      γc              3                                ⁢                      f            2                    ⁢          Δ          ⁢                                          ⁢                      f            .                                              Equation        ⁢                                  ⁢        3            One can see an important result; the number of modes contained in some frequency range is proportional to the volume, and to the square of the frequency. Doubling the volume doubles the number of modes in a frequency range. Double the frequency quadruples the mode structure density. The more modes excited, the better the mode-mixing and the more uniform the microwave exposure inside of the reverberation chamber. In electronic mode stir excitation, one can see that bigger test chambers and higher frequencies are better.
A rule-of-thumb that has often been quoted in electronic mode-stir exposures is that for a given chamber volume and test frequency, the excitation bandwidth should be sufficient to excite at least 60 modes in an over-moded chamber for a statistically significant electric field uniformity. This may arrive at average electric field uniformity variations of about 1 dB. Since more is better, more than a hundred modes should be excited to achieve better than 10% maximum field variation.
Despite the promise of electronic mode-stir to establish rapid electronic testing on electronic time-scales, the concept still has enormous weaknesses. The first weakness is often overlooked, yet very important in RF testing. The typical implementation of electronic mode-stir generation, using a noise generator and a balanced mixer, results in a randomly-generated set of frequencies about the center frequency with the exclusion of the center frequency itself.
The second weakness of traditional electronic mode-stir exposure is far worse. Devices under examination that actually do have significant microwave frequency vulnerabilities will, almost by definition, have correspondingly significant energy absorption at those frequencies. This is a disaster for electronic mode-stir testing. DUEs with significant frequency-dependent microwave absorption therefore change the mode-structure of the excited chamber, and therefore change the average electric field incident upon themselves. Simply put, microwave-absorbing materials or devices exposed in electronic mode-stir chambers change the conditions of their own exposure. In searching for methods of standard industrial microwave exposure, electronic mode-stir is often seen as the first method to discard for this reason alone. There is thus a present need for a method of improving the uniformity of standardized microwave exposures while maximizing the exposure field level for a given RF input power.